From numerical values to user parameters

[uellue]

The model should not only help to apply corrections based on metadata to raw data, but it should also help to transform correction transformations into parameters that users understand.

Example: The model generates a rough correction transformation. The consuming software performs a numerical optimization of the transformation to match a reference dataset, for example a calibration measurement.

The user then wants to obtain an understandable value back from that optimized transformation and optionally save it as a new version of the relevant metadata . Example: "Your defocus was most likely x and not y" --> Attach "correction_abc" to list of available corrections --> apply "correction_abc" to following analyses.